Qué procesos paroxísticos pueden simular una CE?

In this section, a general description of materials and methods used in the studies performed for this thesis is presented. Details regarding procedures can be found in each of the papers. The Uppsala Ethical Committee on Animal Research approved the research protocols and procedures involving the use of animals.

4.1 Study designs (papers I-IV)

Paper I: The aim of this observational study was to register how hooves of show jumping horses approach and impact the ground in jump landings during elite level competition. Digital high-speed video recordings of the landing distal limbs of horses were made during one international Concour de Saut (CSI) 2* and one national elite-level show jumping competition in Sweden during the outdoor season. The competitions were selected as accessible high ranked events in the geographical region around Stockholm during May and June 2009. Landing spots were chosen based on distance to the arena border. The camera was placed safely outside the arena, but was close enough to provide a good field-of-view. The location for registration was also chosen to make sure that the approach and departure from the jump was in a straight line in order to minimize errors due to out of plane movement. The study design allowed no interaction with the subjects. Thus, markerless tracking was performed to study hoof movement in the recorded files from within the calibrated area.

Paper II and III: An experimental study was designed to investigate hoof- surface impact accelerations in horses during canter, jump take-off and landing. Three riders, who rode the same horses throughout the experiment, and five warmblood show jumpers, were recruited from the Equine Studies

program at the Swedish University of Agricultural Sciences. All four hooves of the horses were equipped with two uniaxial accelerometers mounted orthogonally on the lateral hoof wall. Fences were randomly varied for each horse between two types (up-right/oxer) and three heights (0.9-1.3 m, adjusted to the horses’ competition levels). To enable a preliminary investigation of how surface type and level of water content affected hoof-surface impact accelerations, the experiment was repeated in two arenas with one of the two surfaces tested at two levels of water content. An in-situ surface testing device

(OBST) was used to record impact accelerations on the surfaces that could be

compared to corresponding parameters from the horses’ hooves.

Paper IV: The study was designed to describe the construction, material composition and functional properties of show jumping competition and warm- up arenas in highly ranked events (CSI 4-5*) by objective and subjective methods. We also aimed to investigate how the objective, in-situ measurements

of arena functional properties from an OBST, were associated with top-level

riders’ perceptions of these properties. Nine international show jumping events, in six European countries, were selected based on the likelihood of having the same riders participate in several events and for geographical accessibility to facilitate moving the test equipment. Twenty-five competition and warm-up arenas in these events were assessed. A questionnaire was developed for the subjective assessments of the arenas. All riders on the starting lists for the events were asked to evaluate the surfaces subjectively using visual analogue scales.

4.2 Study populations (papers I-IV)

4.2.1 Horses

A summary of information regarding the horses included in studies I-III are presented in table 2. Since hoof-ground interaction was studied in these horses, the available level of information about shoeing is included in the table. In paper IV no horses were directly observed. However riders’ assessments of the surfaces were made after riding one or several horses on the arenas. In this sense the group of ridden horses acted as a mediator for the riders experience of the surfaces’ properties. Details of the horses’ attributes were not compiled, but given the regulations of the competitions in which the study was performed, the age group of the horses was greater than seven years. In order to qualify for these competitions the horses must have been at the top performing level in the international show jumping sport. Mares, geldings and

stallions from different warmblood breeds and Thoroughbred-warmblood crosses are normally represented in this group (Boswell et al. 2011).

Table 2. Summarized data for horses in studies I-III.

Study Number of horses per study Breed Age mean ± standard deviation Sex Competition level Shoeing I 39 European warm- bloods 10.4 ± 2.3 years 18 mares 17 geldings 4 stallions National elite International CSI 2* (1.40-1.45 m fence heights) Information on shoe types not available Screw-in studs were used on the turf arena II and III 5 European

warm- bloods 10.6 ± 2.9 years 2 mares 3 geldings Novice (1.10 m) to Intermediate, (1.30 m) All horses wore regular steel shoes 4.2.2 Arenas

Information about the arenas included in the studies performed for this thesis is summarized in table 3. In total 29 arenas were studied, of which 70% had a sand-fibre top layer, 14% turf, 10% sand and 3% (one arena each) sand- woodchip and waxed sand-fibre. In the temporary arenas (66%) surface material was placed on top of existing floors in indoor sports arenas. Permanent arenas generally use a multi-layered base construction of compacted aggregate.

Table 3. Summarized data for arenas in studies I-IV. Study Arenas per study (n) Indoor/ Outdoor Type of top layer composition Temporary / Permanent Primary use: Training/ Competition/ Warm-up Material specification available I 2 Outdoor Sand Permanent Training and

competition

No Outdoor Turf Permanent Competition No II & III 2 Indoor Sand-Fibre Permanent Training and

competition

Yes, in paper III Indoor Sand-

Woodchip

Permanent Training Yes, in paper III

IV 25 Outdoor Sand Permanent Warm-up Yes

Outdoor Sand Permanent Warm-up Yes Outdoor Turf Permanent Competition No Outdoor Natural turf Permanent Warm-up No Outdoor Natural turf Permanent Warm-up No Outdoor Sand-Fibre Permanent Competition Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Competition Yes Outdoor Sand-Fibre Temporary Competition No (missing) Outdoor Sand-Fibre Temporary Warm-up No (missing) Indoor Sand-Fibre Temporary Warm-up No (missing) Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Competition Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Competition Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Competition Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Warm-up Yes Indoor Sand-Fibre Temporary Competition Yes Indoor Sand-Fibre

with waxed sand below

Temporary Competition Yes

Indoor Waxed sand- Fibre

4.2.3 Riders

The riders (n=3) included in the experiment described in papers II and III were second and third year students at the Equine Studies Program at the Swedish University of Agricultural Sciences in Sweden. All of the riders were female, aged 22-26 years who weighed 52-65 kg and competed at intermediate to advanced level (1.20 to 1.40 m fence heights). Age, weight and gender, and weight were not recorded for the riders in studies I and IV. All of these riders were from the elite ranks for the sport, and in study IV the riders were at the very top international level.

4.3 Kinematic methods (papers I, II, III)

The cinematographic recordings of hoof movements prior to and during surface impact in paper I were performed using a digital high-speed camera (Fastec Imaging, TroubleShooter 1000). The recording rate was 1000 frames/s and the resolution 640 x 480 pixels. The height over ground of the camera aperture was 98 cm and the horizontal distance to the expected centre of landing area 320 cm. The camera was tilted approximately 17° downwards. Calibration was made by filming a folding ruler, giving a vertical and horizontal reference, at three different distances from the camera in the region of interest that provided calibration factors accommodating for hoof landing position (in depth). A software for markerless tracking (Qualisys Video

Analysis, QUALISYS) was used to determine the movements of four points on

each hoof using pattern recognition algorithms. Based on these data, the best fit rigid body transformation were calculated in MatLab (MathWorks Inc.), after application of a fourth order forward-backward Butterworth low-pass filter with a cut-off frequency of 200 Hz. This resulted in 2 translations and one rotation (hoof pitch). Total landing speeds as well as vertical and horizontal components, were calculated as an average from position data over 5 ms pre- impact. The angle between the horizontal plane and the movement path of the landing hooves was determined by calculating a continuous slope (regression) over 6 consecutive data points prior to impact. From the moment of first contact with the ground to the end of hoof braking, the maximal value for vertical deceleration and the maximal value of horizontal deceleration were calculated. The temporal differences between these peaks were described. The acceleration data from hoof-surface impacts in paper II and III were produced using two single axis ± 250 g accelerometers (ADXL193, Analog Devices) attached to the hoof by a metal fixture (total weight of 22 g). The accelerometers were placed orthogonally in vertical and fore-aft direction.

Signal wires from the accelerometers were plugged into a 14-bit data logger

(DataLOG MWX8, Biometrics) which was carried by the riders in a waist bag.

Sampling rate was set to 1000 Hz. From the collected data, 15 strides around

the jumps were selected with a custom written MatLab script.The signal was

filtered with a fourth order forward–backward low-pass Butterworth filter with a cut-off frequency of 400 Hz. In the impact complex at the beginning of the stance, the peak vertical deceleration was identified. The range between maximal horizontal deceleration and acceleration was used as a measure of the magnitude of horizontal ground interaction during impact. The quotient of the acceleration vectors was calculated in order to describe the relation between the horizontal and vertical ground interaction in the early hoof-surface interface. Break duration was calculated as time in milliseconds from the first vertical deceleration peak to hoof standstill. Each impact was classified as leading or trailing forelimb or hind limb respectively and stride types (canter, jump take-off or landing) was assigned.

4.4 In-situ measurements of functional properties using the Orono Biomechanical Surface Tester (papers III and IV)

The OBST was used for objective, standardized surface assessment in paper III

and IV. The device interacts with the surface in both vertical and horizontal directions. A metal hoof connected to a heavy mass, guided by angled rails, was dropped on to the surface (see supplementary material to paper IV). As the hoof impacts the ground the falling mass above transfers additional load onto the hoof by a shorter vertical axis compressing the spring and damper, at the same time allowing a forward slide of the hoof. In paper III the device’s long guiding rails were positioned in a more acute angle to the vertical compared to the settings used in paper IV. In table 4 modifications of the original design by Peterson et al. (2008a) of the test device is presented. Tri-axial accelerations, tri-axial loads and position data were acquired from the device through nine channels of data recorded with 16-bit resolution at 5000 Hz using a custom written MatLab data acquisition and analysis script. In paper III only the peak vertical deceleration of the metal hoof was used.

Parameters derived from the sensor outputs from the OBST were used to

measure functional properties of the surfaces (see descriptions in table 5). Each measurement was chosen based on its appropriateness to define the biomechanics of the property. Impact firmness was characterized by the peak vertical deceleration of the metal hoof at impact, aimed to represent the shock experienced by the horse at hoof impact. Cushioning, describing the surfaces ability to absorb and reduce peak force, was determined using the peak vertical

force from the tri-axial load cell. Grip was represented by the amount of forward slide of the metal hoof on the surface during loading. Responsiveness relates to the deformation and elastic recovery of the surface and was measured as a quotient of the compression and recoil time of the spring-mass-damper system. Uniformity, representing the spatial variation over the arena, was calculated by taking the ensemble mean of the coefficients of variation (CV), defined as ‘the standard deviation divided by the mean’, for each functional property; impact firmness, cushioning, grip and responsiveness of that arena. For detailed descriptions of signal processing, parameter calculations and graphical representation of the signals in the time domain see Supplementary material to Paper IV and additional results in section 4.8.

Table 4. Adjustments to the Orono Biomechanical Surface Tester compared to the original design

described in Peterson et al. (2008a)

Settings/design Peterson et al. 2008a Paper III Paper IV Hoof and shoe Hoof cast from a two part

casting rubber (Duo-Matrix Neo, Smooth-On, Easton, PA, USA)

Metal hoof with a standard iron shoe size 2

Metal hoof with a standard iron shoe size 2 Drop height (vertical) 1.83 m 0.84 m 0.84 m Weight of falling mass 30 kg 33 kg 33 kg Impact energy 540 J 272 J 272 J

Angle of long rails

(from vertical) 12° 8° 12°

Angle of short rails

(from vertical) 7° 0° (hoof lands flat) 0° (hoof lands flat) Spring + damper Gas spring (EFA 20- 50-

FC, Efdyn, Tulsa, OK, USA) Metal spring (Ashfield Springs Ltd. s421) + Industrial damper (Enidine OEM 2.0Mx4CMS 100mm, setting #2) Metal spring (Ashfield Springs Ltd. s421) + Industrial damper (Enidine OEM 2.0Mx4CMS 100mm, setting #2)

4.5 Laboratory material tests (papers III and IV)

In order to characterise the arena surface materials of the top layers a material sample of approximately 1 kg was collected from the surfaces investigated in Papers II, III and Paper IV. The turf arenas in Paper IV could not be sampled and from one event in this study the collected material was misplaced (marked as missing, see table 3). Particle size distribution was determined by sieving

and sedimentation, water contentwas measured by drying samples at 45°C to a

constant mass, the percentage of organic content was determined by burning off the organic materials from an oven dried sample in a furnace, and when applicable, the wax content was registered using the Soxhlet extraction method. In paper III a bulk density test was also performed which describes how the material compacts under different moisture conditions.

4.6 Questionnaire (paper IV)

In study IV a questionnaire was developed to record riders’ assessments of functional properties of the arenas at the show jumping events. The properties and the words describing them were selected on the basis of being of biomechanical relevance to the horse, familiar to the riders and also possible to measure mechanically. One questionnaire, per event and rider, was used to evaluate properties and overall scores for each arena by using visual analogue scales. Short descriptions of the properties and the verbal anchors of contrasting adjectives used at the end-points of each scale are presented in table 5. The visual-analogue scores were measured 0-100 ordinal scale and then transformed to a 0 to 5-rating, with the resolution unchanged. Written definitions and in depth explanations in English of each of the functional properties (according to Hobbs et al., 2014 pp. 20-21) were given to the riders the first time they were approached with the questionnaire to facilitate interpretation. The riders were provided a translated version of the explanation and questionnaire in French or German if they wished. The riders were asked to evaluate the arenas in comparison to arenas on other events of CSI 3* or higher ranking, that they had participated in during the last five years. Riders assessed surfaces after they had ridden at least one horse on all included surfaces at the event.

Table 5. Functional properties of arena surfaces used in the questionnaire. The verbal anchors

describe end-values for each property and the short version of description was given to each visual-analogue scale

Functional property ‘High-end’ verbal anchor

‘Low-end’ verbal anchor

Short description

Impact firmness Hard Soft The shock experienced by the horse and rider _{when the hoof contacts the surface.} Cushioning Deep Compacted How much a surface is supportive compared _{to how much it gives when riding on it.}

Grip High grip Slippery How much the horse’s hoof slides during _{landing, turning and pushing off.} Responsiveness Active Dead How active or springy the surface feels to the _rider.

Uniformity Uniform Variable How regular the surface feels when the horse _{moves across it.} Consistency No change Changeable How much the surface changes with time and _use.

4.7 Statistical methods (papers I-IV)

The statistical analyses in this thesis were performed using the statistical software SAS (SAS Institute Inc., USA). Mixed models were used for data acquired in all four papers. Several fixed effects and their interactions were studied using these models and the hierarchical structures of the data and repeated observations (non-independence of data points) was accounted for by adding random effects (horses, riders, events or their combinations). In Paper III a Student’s T-test was also used to investigate the difference in measured

impact deceleration from the OBST on the different surfaces. Normality of the

distributions of the outcome variables was always tested as described in the papers.

4.8 Additional OBST data from training and competition arenas

As a reference material to the arena measurements made with the OBST,

presented in paper III and IV, additional data from a selection of competition, warm-up and training surfaces are presented here. Descriptive statistics can be found in table 6 and examples of signal traces in the time domain are given in figures 2-5. The signals are chosen to display both typical characteristics and more extreme signal patterns found in measurements from these arenas.

Measurements were performed from 2012 to 2015. The OBST settings were

identical to those in paper IV. Arena attributes have not been specified since these data are provided to enable a general comparison that highlights the

between and within arena variation in measured values. Sample points (drop places) within the arenas were spaced approximately in a 15-20 m grid, resulting in 9-12 drop places in an average-sized indoor training arena. Data are presented for 94 measurement occasions on competition and warm-up arenas on 4 and 5* FEI competitions in Sweden and Germany and for 428 training arenas in Sweden and UK. Each arena can have been measured at several occasions (different days and different preparations). For competition arenas (n=94) there were 56 unique arenas, the mean number of times an arena was measured were 1.68 and the median was 1. For training arenas (n=428) there were 164 unique arenas, the mean number of times an arena was measured were 2.48 and the median was 2.

Table 6. Descriptive statistics of data from measurements performed with the OBST from 2012 to

2015. Means over arena-mean values are presented for vertical peak load, peak load rate and acceleration. The coefficient of variation (CV) for the parameters was computed for each arena and then averaged for the arena group presented (competition/warm-up or training) to indicate mean within-arena variability.

Mean Mean within arena CV Median Standard deviation 5th percentile 95th percentile Competition & warm-up arenas measurements n=94

Peak vertical load (kN) 14.7 12% 14.9 1.5 12.0 16.9 Peak vertical loading rate (kN/s) 4670 33% 4284 1655 2934 8517 Peak vertical acceleration (g) 87 28% 86 19 62 126

Training arena measurements n=428 Peak vertical load

(kN) 11.7 20% 11.7 3.6 7.2 17.4 Peak vertical loading rate (kN/s) 3000 57% 2779 1986 1194 5963 Peak vertical acceleration (g) 75 94% 73 30 16 123

Figure 2. To the upper left vertical acceleration, registered on the metal hoof is presented in g. In

the upper right panel the vertical load from the triaxial load cell is displayed in kN. The lower left panel shows the horizontal load in kN and to the right vertical loading rate in kN/s is presented.